J . Am. Chem. SOC.1983, 105, 7267-7271
7267
Nucleophilic Reactions of F3C- at sp2 and sp3 Carbon in the Gas Phase. Characterization of Carbonyl Addition Adducts Richard N. McDonald* and A. Kasem Chowdhury Contribution from the Department of Chemistry, Kansas State University, Manhattan, Kansas 66506. Received April 22, 1983
Abstract: The reactions of F3C- with CH3Br and CH3CI established the medium kinetic nucleophilicity of F3C- on Bohme's reactivity scale for gas-phase SN2reactions. The reactions of F3C- with (CH3)2C=0and CH,CO2CH3proceeded by competitive bimolecular H+ transfer and termolecular carbonyl addition giving the corresponding adduct anions m / z 127 and 143, respectively. F 3 C reacted with the esters C6H5C02CH3,CF3C02CH3,and (CH30)2C=0 both by SN2displacement forming the corresponding carboxylate anions and by carbonyl addition yielding the adduct anions; with CF3C02C2H5and CF,CO2C(CH,),, the competitive bimolecular reaction channel involved E2 elimination giving CF,CO2-. The major reaction channel of F3C- with HC02CH, was the Riveros reaction that produced the series of cluster ions F3C-(HOCH3), F3C-(HOCH3)2,CH30-(HOCH3),and CH30-(HOCH3)2,along with a minor amount of carbonyl addition. The fast termolecular reaction of F,C- with (CF,),C=O exclusively formed the adduct (CF3)3CO-( m / z 235) which was characterized as the bound, tetrahedral structure by bracketing its proton affinity. The reaction of F3C-with C 0 2giving CF,CO< was established as a termolecular process when the "apparent" bimolecular rate constant was shown to be PHsdependent. These results demonstrate unequivocallythat the reactions of gas-phase nucleophiles with the carbonyl group of ketones and esters proceed by addition yielding the corresponding adduct oxyanions which is analogous to the related processes in the condensed phase.
The trifluoromethyl group has been used for years in condensed phase reactions as a "chemically inert" group to destabilize carbonium ions (e.g., u + =~ 0.520 and u + = ~ 0.612 are substituent constants for F3C)' and stabilize carbanions. In the gas phase, similar effects of the F3C group are observed in the acidity (Azfoacid) of ring substituted phenols2 and anilines.3 This effect carries over to the series of alcohols C H 3 0 H (AGOacid= 372.6 f 2 kcal CF,CH,OH (AGOabd= 356.8 f 2 kcal mol-'),, (CF,),CHOH (AGOabd= 340.8 f 2 kcal mol-')? and (CF,),COH (AGOacid = 326.7 f 2 kcal m01-l)~where the CF, groups acid strengthening effect is additive (6AGoacid 15 kcal mol-'). Our interest in the chemistry of F3C- stems from our studies of 1,2- vs. 1,4-addition reactions of nucleophiles with +unsaturated moleculesSand the observation that F3C- served as a good nucleophilic initiator for the anionic oligomerization of H2C=CHC02CH3.6 These factors coupled with our interest in measuring absolute (and relative) reactivities of nucleophilic addition to simple carbonyl-containing molecules' suggested that if tetrahedral adducts (1) are indeed produced in the gas phase according to eq 1, Nu:- = F,C- should markedly increase the
-
Nu:-
+
R,C=O
i"
T- NU-C-R
i)
I R
1
probability for the direct observation of adducts 1. This expectation was supported by the observation of the anion m / z 197 of the correct mass for the adduct (CF3)2C(O-)OCH3in the reaction of allyl anion (C3Hs-) with CF,CO2CH3 by the sequence in eq ii-iv.* The reasonably large proton affinity of F,C- (PA = 375.6 ~
~~~
(1) Brown, H. C.; Okamoto,
Y. J.Am. Chem. SOC.1958,80,4979-4987. (2) Fujio, M.; McIver, R. T.; Taft, R. W. J. Am. Chem. SOC.1981, 103, 4017-4029. (3) (a) Bartmess, J . E.; McIver, R. T. In 'Gas Phase Ion Chemistry"; Bowers, M. T., Ed.; Academic Press: New York, 1979; Voi. 2, Chapter 1 1 . (b) Bartmess, J. E.; Scott, J. A.; McIver, R. T. J.Am. Chem. SOC.1979, 101, 6046-6056, 6056-6063. (4) Koppel, T.; Taft, R. W., unpublished results; private communication from Professor Taft. ( 5 ) McDonald, R. N.; Chowdhury, A. K. J. Phys. Chem. 1982, 86, 3641-3645. (6) McDonald, R. N.; Chowdhury, A. K. J. Am. Chem. SOC.1983, 105, 2194-2203. (7) McDonald, R. N.; Chowdhury, A. K. J. Am. Chem. Soc. 1983, 105, 198-207.
C3H5-
+ CF,CO,CH,
-+
[CF3COT]* + C H ~ C H ~ C H Z C H ~ (ii)
[CF3COY]*
F3C-
-+
F3C- + CF3C02CH3
He
,
~
(CF3)2C(O-)OCH3 ( m f z 197)
(iii) (iv)
f 2 kcal is only slightly less than that of C,HsO- and indicates that exothermic H+ transfer from a number of neutral reactants will be a competing reaction channel with addition to the carbonyl group. Bohme et reported that stable addition adducts were produced in the gas-phase reactions of H-, HO-, and C H 3 0 - with H 2 C = 0 following third-body collisional stabilization with the buffer gas of the initially formed excited adduct anions. That adducts such as 1 might not be intermediates in reactions of nucleophiles with various derivatives of carboxylic acids gained precedence in the recent report by Kim and Caserio.'O These latter authors examined the gas-phase acyl transfer from conjugate acids of carboxylic acids and certain derivatives to neutral nucleophiles and concluded that direct displacement (SN2type) occurred rather than addition followed by elimination. DeTar" has reconsidered the data for several acyl transfer reactions in the condensed phase and suggested that certain of these processes may also proceed by a direct displacement mechanism.12 Therefore, it appeared essential to observe and characterize a series of adducts 1 formed in reactions of nucleophilic anions with carbonyl compounds of varying structural types.
Experimental Section The flowing afterglow (FA) apparatus with a modular flow tube (Figure 1) has been previously described.6 Briefly, the ion of interest is prepared in the upstream end of a stainless steel flow tube (140 X 7.15 cm i.d.) by electron impact with small concentrations of neutral molecules or by fast ion-molecule reactions using previously generated ions. In (8) McDonald, R. N.; Chowdhury, A. K. J. Am. Chem. SOC.1982, 104, 901-902. (9) Bohme, D. K.; Mackay, G. I.; Tanner, S . D. J.Am. Chem. SOC.1980, 102, 407-409. (10) Kim, J. K.; Caserio, M. C. J. Am. Chem. Soc. 1981,103,2124-2127. (1 1) DeTar, D. F. J. Am. Chem. SOC.1982, 104, 7205-7212. (12) We wish to point out that the gas-phase result reported by Fukuda and McIver (J.Am. Chem. SOC.,1979, 101, 2498-2499) of CH,CO,Ph + -OCH, CH3C02-[+ CH30Ph] was in error. The correct product was PhO- produced by carbonyl addition followed by anionic fragmentation; see: Kleingeld, J. C.; Nibbering, N. M. M.; Grabowski, J. J.; DePuy, C. H.; Fukuda, E. K.; McIver, R. T. Tetrahedron Lett. 1982, 23, 4755-4758.
-
0002-7863183 /1505-7267%01.50/0 , . 0 1983 American Chemical Society I
+ C02
1268 J. Am. Chem. SOC., Vol. 105, No. 25, 1983
McDonald and Chowdhury
Table I. Summary of Kinetic and Product Data for the Ion-Molecule Reactions of F3C-
rxn. 110.
1 2 3a 3b 4 5a 5b 6a 6b 7a 7b 8a 8b 8c 9a 9b 1Oa 10b 1l a llb 12
ion
+ neutral reactants + CH,Br +
F,CF,C-
+ CH,CI + (CH,),C=O + (CF,),C=O + He + CH,CO,CH,
F,C-
+ C,H,CO,CH,
F,C
+ CF,CO,CH,
T,C-
+ CF;,CO,C,H,
r,c- + CF,CO,C(CH,),
r,c- + (CH,O),C=O F,C-
+ HCO,CH,
r,c- + CO, + He
fraction of product ion iignal
products [assumed neutral]
Br- [ + F,CCH,]
-+ Cl- [ + F,CCH,] -+ CH,=C(O-)CH, [ + HCF,]
,
-+ K H , ),C(O-)CI: + (CI:,),CO- + He + CH,=C(O-)OCH, [ + HCI’,] + CH (CF )C (O-)OCH -+ C, H (CF ,)C(O-)OCH, -+ C,H,CO; [ + F,CCH,] -+ CF,CO; [ + F,CCH,] -+ (CT:,1, C(O-)OCH, + (CF,),C(O-)OC,H, -+ CF,CO; [ + F,CC,H,] + CF,CO; [ + C,H, HCF,] + (Cf~,),C(O~)OC(CH,), -+ cr:,co,- [ + C,H, + HCF,] + ( C H 0) C (0.)C F + CH,OCO,- [ + F,CCH,l -+ l-,C-(HOCH,) + F,3C-(HOCH,), CH,O-(HOCH,) CH,O-(HOCH,), -+ C I: CH (0‘)OCH -+ CF,CO; He
, , ,
+
,
, ,
,
,
+
+
+
1.00 1.00 0.85 0.15 1.00 0.72 0.28 0.77 0.23 0.67 0.33 0.90
0.10 0.92 0.08 0.92 0.08 0.96
kADO,‘ cm3 kcal mol-’
ktotal>bcnl3 molecule-’ s-’
-62.5 -54.6 -6.8
( 2 . 1 0.2) x 10-10 (5.6 t 0.1) x l o - ” (2.5 0.1) x 10.”
1.4 x 10-9 1.6 x 10-9 2.1 x 10-9
-25.1 -66.9 -4.6 -10.0
(4.5 t 0.3) x (5.5 I O . 1 ) x l o - ”
8.1 X
*
+_
molecule-’
s- I
1.5 x 10-9
e -57.8 -70.1 -33.7 -33.7 -60.0 -37.1 -33.7
x 10-9
(1.1 t 0.1) x 10-9
1.7
(9.1
1.8 x 10-9
+_
0.1) X 10.’’
(8.6 I 0.1) x IO-”
-35.0
(2.9 t 0.3) x lo-“ (4.2 t 0.5) x l o - ”
1.5 x 10-9
1.1 i:
4.4 x
0.04
1.00
-38.8
d
Apparent bimolecular ktota!’s (see Result5 section for discussion) are estimated to be accurate to +25%. Errors a See text and ref 15. Calculated collision rate constants using the average dipole orientation theory, given are standard deviations from multiple determinations. Termolecular ktotal in units of cm6 sC1 calculated by dividing the “apparent” bimolecular rate constant for decay of ref 16. r , C - by [ H e ] ; [He] = 1.6 X 10l6atoms cin-, a t 0.5 torr. e Not measured. reagents were distilled just prior to use, and a center-cut, constant boiling sample was transferred to a gas storage bulb after three freeze-pumpthaw degassing cycles. The gas reagents were used directly.
6
DFidSlON -.UP
Figure 1. Diagram of the flowing afterglow apparatus. these experiments, flowing a mixture of H 3 N in the buffer gas, helium, over the electron gun yields H2N- by dissociative electron attachment (thermal or near-thermalized 0 ) Addition . of HCF, to this flow via a port IO cm downstream of the electron gun yielded vibrationally excited F3C- ions by H + transfer to H2N- ( k = (1.3 f 0.2) X cm3 molecule-’ s-l). The vibrationally excited F3C- ions were cooled to their vibrational ground state by numerous collisions with the buffer gas ( P H =~ 0.5 torr, 0 = 80 m SKI) in the next 30 c m of the flow tube prior to reaching the neutral reactant inlet port for addition of the carbonyl-containing substrates. The length of the ion-molecule reaction region was 62.5 cm. The flow was exhausted by means of a large pumping system, but it was sampled into a compartment containing a quadrupole mass spectrometer operating at -IO-’ torr which monitors the ion composition of the flow. Kinetic measurements are straightforward and were carried out under pseudo-first-order conditions with the neutral reactant in large excess compared to the ion concentration. The ion-molecular reaction distance was held constant while the concentration of neutral reactant was varied. From the slope of the semilog plot of log [F3C-] vs. neutral reactant, the rate constant for the bimolecular ion-molecule reaction was c a l c ~ l a t e d . ~ ~ The estimated error in the rate constants is f25%. The product ions were also monitored during these experiments which yield the branching fractions reported. Helium (99.99%; Welders Products) was the buffer gas in these experiments. It was further purified by passage through two traps filled with Davison 4A molecular sieves cooled with liquid N2 and then warmed to 298 K in a glass coil prior to introduction into the upstream end of the flow tube (Figure 1). Gas and liquid reagents were available from commercial sources (Matheson, PCR, Fisher, Aldrich). The liquid (13) McDonald, R. N.; Chowdhury, A. K.; Setser, D. W. J . Am. Chem. SOC. 1980, 102, 6491-6498.
Results The kinetic and product data for the ion-molecule reactions of F3C- are summarized in Table I. In all cases, clean pseudofirst-order decay plots of the log ion signal of F3C- ( m / z 69) vs. concentration of the neutral reactants were observed. The recovery of product ions was always >90% of the decay of the F3C- signal. The product channel branching fractions in Table I are the relative product ion signals observed from the mass spectrometer. In reactions 3 and 5-1 1 in Table I, the rate constants for decay of the F3C- ion signal are listed as bimolecular rate constants. While this is true for the HC transfer and sN2 displacement product channels, observation of the adduct anions requires third-body (termolecular) collisional stabilization of the vibrationally excited adducts by the buffer gas. Since these reactions showed no P,, dependence (varied from 0.4 to 1.1 torr) on the “apparent” bimolecular rate constant for decay of F3C- or on the branching fraction, saturated termolecularity must be present in the adduct forming product channels even at our lowest buffer gas pressure. Discussion Reactions of F3C- with CH3Br and CH3CI. For the kinetic nucleophilicity of F3C- toward saturated sp3 carbon to be determined and placed on the reactivity scale of the sN2 displacement reactions of Bohme et al.,14athe reactions of F3C- with CH3Br (reaction 1) and CH3C1 (reaction 2) were examined. In both of these reasonably exothermic15 reactions, formation of the corre(14) (a) Tanaka, K.; Mackay, G. I.; Payzant, J. D.; Bohme, D. K. Can. J . Chem. 1976, 54, 1643-1659. (b) See also: Olmstead, W. M.; Brauman, J. I. J . Am. Chem. SOC.1977, 99,4219-4228. (1 5 ) The thermochemical data for calculation of AHO’s come from
standard sources: (a) Cox, J. D.; Pilcher, G. ‘Thermochemistry of Organic and Organometallic Compounds”; Academic Press: New York, 1970. (b) Benson, S. W. “Thermochemical Kinetics”; 2nd ed.;Wiley: New York, 1976. (c) Reference 3a. (d) Calculated AHlo’s for some neutrals were carried out with use of the additivity group tables in ref 15b. ( e ) W,O(Br-) = -50.8 kcal mol-’. (0AHio(C1-) = -54.4 kcal mol-]. ( 9 ) AHIo(F,C-) = -157.2 kcal m01-1.3a
J . Am. Chem, Soc., Vol. 105, KO. 25, 1983 7269
Nucleophilic Reactions of F 3 C sponding halide ion was observed. However, the rate constants for both reactions are significantly less than the calculated collision limits, kADo.16 This requires that modest barriers separate the collision complexes (F3C-/CH3X) and (X-/CF3CH3) on these double-minimum potential surface^.'^^^" Thus, the kinetic nucleophilicity of F3C- toward SN2 displacement with CH3X molecules must be judged to be medium.14a Reactions of F 3 Cwith Ketones. The H+ transfer between F3Cand (CH3),C=0 is exothermic by 6.8 kcal mol-I. Therefore, it was no surprise to find that this was the major product channel in their ion-molecule reaction (eq 3). However, it was gratifying
endothermic (225 kcal mol-I) and would not occur under these gas-phase conditions. However, observation of adduct anions in these reactions would establish that addition to the ester carbonyl group is the lower energy pathway of these two processes. The reaction of F3C- with CH3C02CH, (reaction 5) proceeded with a modest rate constant via two product-forming channels, reaction 5a (78%) involved H+ transfer (AHoaCid(CH3CO2CH3) = 371 kcal m01-I)~and reaction 5b (28%) yielded the adduct m l z
+
nCF3
(Sa)
- 10
:I
H,E--'c-cH,
+
17-
HCF,
(3a)
(mlz 5 7 )
F,C-
4- (CH,),C=O
0.15
H3C-c-cH3
I
(3b)
CF3
( m / z 127) I
to observe the minor formation of the ion m l z 127 which was the sum of the masses of the two reactants. The ion m / z 127 was shown not to be the simple hydrogen bonded cluster ion of m / z 57 and HCF3 (H2C=C(O-)CH3. HCF3) since the relative intensities of m l z 57 and 127 in reaction 3 did not change as a function of varying the [HCFJ used to generate the F3C- anion. The reaction of F 3 C with (CF3)*C=0 was examined since only the addition reaction channel was possible. As expected from the reaction exothermicityis and the polarizability of this ketone, this fast reaction occurred with the exclusive formation of the adduct anion m l z 235. That the rate constant for this reaction was 56%
-
FjC-
+ (CFj),C=O
He
(CF3)3C-O( m / z 235)
(4)
of the termolecular collision limit (kAD0)I6indicated the stability/lifetime of the initially formed excited adduct ((CF3),CO-)* and the relative efficiency of collisional stabilization of such excited species in the FA apparatus. The structures of the adducts m / z 127 (eq 3b) and 235 (eq 4), and of the other adducts to be discussed, provide a large density of states in the vibrational and rotational manifold of the adduct. The excess energy of the adduct-forming reaction channel ( A H o ) will be. spread over these many states which reduces the probability of sufficient energy concentrating in the stretching vibration of the C-C bond necessary for retro-addition. This increases the lifetime of the excited adduct which allows for collisional deactivation and observation of the adduct anion. That the above density of states argument cannot be the whole story is seen in the fact that the termolecular addition reactions of H- and HO- with H,C=O occurred at about 3% of the collision limit9 while that of addition of F3C- with (CH3)&=0 was about 0.2% of its collision limit. This suggests that there are also kinetic barriers in these nucleophilic addition processes to carbonyl groups which separate the energy minima of the loose ion-dipole collision complexes and their adducts. Reactions of F 3 Cwith Esters. The reactions of F3C- with esters RC02R' cannot involve direct nucleophilic displacement to yield R'O- and RC(=O)CF3 since such reactions would be strongly (16) Calculated by using the average dipole orientation theory (ADO): Su, T.; Bowers, M . T. J . Chem. Phys., 1973,58, 3027-3037; In&.J. Mass Spectrom. Ion Phvs. 1973. 12. 347-356. (17) Pellehte, M.'J.; 'Brauman, J. I. J . Am. Chem. SOC.1980, 102, 5993-5999. (18) AHfO((CF,),CO-) = -560.3 kcal mol-' (calcd)'Sb,dusing AHoacid((CF,),COH) = 334.3 kcal m01-I;'~ AHf0((CF3),C=O) = -336.2 kcal mol-' ( c a l ~ d ) ' using ~ ~ , ~the group equivalent value of C(F),(CO) = -152.4 kcal
CF,
&=
+35.3 k c d 11101.'
143. The anionic fragmentation of m / z 143 shown in eq 5c ia strongly endothermicZoand could not occur since channel 5b was only exothermic by I O kcal mol-' (Table I ) . The products of reaction 5c would be expected to undergo H+ transfer in their loose ion-dipole complex and yield the enolate anion H2C=C(O-)CF, ( m l z 111) and C H 3 0 H . Anion mlz 11 1 was not observed as a product in this reaction. Further, separate studies have shown that enolate anion m / z 111 does not form a cluster ion w i t h C H 3 0 H which eliminates this cluster as a possible structure for the adduct m l z 143. The reaction of F3C- with C6HSC02CH3was carried out by adding the vapors of the ester through a special high-boiling inlet connected directly to a reservoir of the liquid ester. Although the rate constant for this reaction could not be measured by using this approach, two product-forming channels were readily observed (eq 6). The major process (77%) involved adduct formation ( m / z 0-
I
L C,H,CO,-
i6a)
+
F3CCH,
(61))
( m / z 121)
205) while the minor channel (23%) gave anion m l z 121 (C6HSCOT),the latter presumably formed by Sx2 displacement at methyl c of the ester. Since C6H,C02-should be a fairly good anionic leaving group in the SN2mechanism and AHo (reaction 6b) = -57.8 kcal mol-', this result indicates that addition to the carbonyl group of the ester by F3C- must be a reaction with a low kinetic barrier separating the initial loose iondipole collision complex and the adduct m / z 205. The A H o for reaction 6a is considerably less than that for reaction 6b. We believe that this failure to follow the product channel of greatest exothermicity demonstrates that closer approach of the anionic nucleophile to the ester group, compared to their separation in the ion -dipole collision complex, is energetically favored along the molecular
(20) (a) AHf"(CH,(CF3)C(O-)OCH,) = -265.2 kcal niul-' (calcd)lSbd assuming AHo,,,d(CH,(CF,)C(OH)OCH3) = 363 kcal mol '.*Ob (b) AH',,,, of the hemiacetal CF,CH(OH)OCH, was bracketed between that of HCCI, (AHoacld= 362 f 6 kcal mol-')' and that of CF,CH,OH (AH",. = 364.4 2 kcal mol-').3 We assign the acid strengthening effect of C',.-OCH, in mol-I. hemiacetals and heniiketals to be 1 kcal mol-'. (19) Using AG0,,,,((CF3)3COH) = 326.7 f 2 kcal mol-' and Soacld-(21) AHfo(C,H,(CF,)C(O-)OCH,) = - 242.4 kcal mu1 ' (ialcd)''b,dah((CF,),COH) = SD,,,d(CF3CH20H) = 25.5 cal mol-l deg-'.3a surning AHo,,,d(C6H,(CF3)C(OH)OCH,) = 353.6 kcal iiiul ' ' .Ob
*
1270 J. Am. Chem. SOC., Vol. 105, No. 25, 1983
McDonald and Chowdhury
dipole (approximately along the 6+C=O") rather than elsewhere in the molecule. The reaction of F,C- with CF3COZCH3sets the stage for competitive carbonyl addition and the SN2reaction at methyl C with a "super" anionic leaving g r o ~ p . ~Two , ' ~ ~product anions, m / z 113 (CF3C02-) and 197 ((CF,),C(O-)OCH,), were formed in a ratio of 2/1 (reaction 7) in this reaction which occurred on almost every collision. Since the direct displacement of C H 3 0 -
{(yi)
+ F3CCH3
F3C-
+
(7a)
CF3COzCH3
(7b)
(CF,)ZCOCH,
(miz 197) (+ (CF3)zC=O) from the reactants in reaction 7 is endothermic by 25.7 kcal mol-', the structure of m / z 197 cannot reasonably by represented by the ion-dipole complex (CH,O-/(CF,),CO). The fact that reaction channel 7a is exothermic by 70.1 kcal mol-' 22 means that we may be observing only a fraction of the true amount of the SN2channel. This is the result of competitive decomposition of some of the vibrationally excited product CF3C0; ions giving F3C- COz (W = +38.8 kcal mol-') and collisional stabilization by the buffer gas yielding the observed CF,CO2- anions.s Therefore, the ratio of reaction channels (7a/7b) = 2 is a lower limit. Since k,,,] is 65% of the calculated collision limit, ICADO, the additional complication of this decomposition pathway means that reaction 7 occurs even closer to the collision limit. It is known from condensed phase studies of the sN2 displacement reaction that nucleophilic displacement at c, of C2H5X derivatives is kinetically slower than at methyl C of CH3X deH s30.z3 This factor rivatives leading to an average k C H 3 / k C 2= appears to carry over to the very fast reaction of F3C- with CF,COZCzHS(reaction 8) which occurred at 50% of the collision limit. Here we observed that formation of the adduct anion m / z
+
0-
I
F3C-
+
CF3CO2C2H5
€
(83)
(CF3),d0C,H5
(miz 211) CF3C02-
t F,CC&IH,
(m/z 113) f
CH ,,
+
or
(8b)
HCF,
(8~)
21 1 was 90% of the product channels with only 10% leading to production of CF3C02-( m / z 113). The minor amount of CF3C0,- in reaction 8 may arise by two separate mechanisms, SN2 nucleophilic displacement at ethyl C, of the ester (reaction 8b; AHo = -60.0 kcal mol-]) or E2 elimination (reaction 8c; AHo = -37.1 kcal mol-') since both reactions are exothermkZ4 CF,COz- is not sufficiently energetic to decompose to F3C- and
coz.
In an attempt to distinguish between the SN2(reaction 8b) and E2 (reaction 8c) mechanistic possibilities, the reaction of F3Cwith CF3C02C(CH3),was examined (reaction 9). Since sN2 displacement at C, of the tert-butyl is not possible, only elimination reaction 9b could yield C F 3 C 0 y . That the ktotal'sand product channels of reactions 8 and 9 are the same within experimental error lead us to conclude that the reaction channel yielding CF3CO; in reaction 8 is primarily or exclusively reaction 8c, the E2 elimination pathway. Thus, k C H 3 / k C I> H S7 (0.67/0.10) is established in these gasphase SN2reactions which occur a t or near the collision limit.25 (22) (a) AHfo(CF3C0,CH3) = -240.7 kcal mo1-I ( c a l ~ d ) . l ~ ~ , (b) ~,'* AH,o((CF3)2C(O-)OCH3) = -431.6 kcal mol-' ( ~ a l c d ) ' ~assuming ~,~ Mnae,d((CF3)2C(OH)O~H3) = 347.4 kcal (23) Streitwieser, A. Solvolytic Displacement Reactions"; McGraw-Hill: New York, 1962; pp 11-13. (24) (a) M10(CFIC02C2H,) = -248.9 kcal molP ( c a l ~ d ) . ~(b) ~~.~ AHP((CF3)2C(O-)OC2H5)= -439.8 kcal mo1-I ( ~ a l c d ) "assuming ~~ the same Noacld of this hemiketal as in ref 22b.
We believe that in each of these reactions, a single loose collision complex [F3C-/CF3CO2R]is produced which then is partitioned between (a) addition to the ester carbonyl group giving the adduct and (b) attack at the alkyl group (SN2 for R = CH, or E2 for R = CzHs) producing CF,COT. The barriers separating the collision complex and the next stage of the product channels must be small to account for the very fast rates of these reactions. We do not know the degree of enthalpic and/or entropic contributions H 5 but determination of the temperature to the k C H 3 / k C 2ratio, effects on these and related SN2 reactions is clearly required. The reaction of F3C- with dimethyl carbonate ((CH30)2C--U, reaction 10) proceeded with a modest rate constant. The major product-forming channel was formation of the adduct m / z 159 along with a minor amount of m / z 75 believed to be the methyl carbonate anion produced by sN2 displacement at methyl c. These results suggest that a partial list of carboxylate leaving group abilities in SN2reactions is CH,COT < C H 3 0 C 0 2 -< C6HSC02< CF3C0y. The order of the first and last two carboxylates follow their known PA'S.^",^^^ The major product channel(s) in the reaction of F3C- with H C 0 2 C H 3involved formation of a series of cluster ions, m / z 101 (F,C-..HOCH,), 133 (F3C-..(HOCH3)2), 63 (CH,O-.. HOCH3), and 95 (CH30-. (HOCH3)z),and a minor amount (4%) of the adduct m / z 129 (CF3CH(O-)OCH3) (reaction 11). The first-formed cluster ion m / z 101 was readily explained via a Riveros reactionz6 shown in eq 13.27 Since the rate of the reaction of F3C- with H C 0 2 C H 3 was slow, considerable HC02CH3 (>1Ol2 molecules ~ m - was ~ ) required to decay the signal F3C- HCO2CH3 --c F3C-e. HOCH3 C O (1 3)
-
+
+
( m / z 101)
of F,C-. This allows secondary Riveros reactions of [base-.. HOCH3] clusters with HCOZCH, to occur.' H+ transfer from CH,OH to F,C- is mildly endothermic (3.6 kcal mol-]).," The similar values of the acid strengths of HCF3 (AHoaad= 375.6 f 2 kcal mol-')," and CH,OH (AHoaad = 379.2 f 2 kcal should produce a reasonably strong hydrogen bonded cluster ion.28 However, the intensity of the ion signal of m / z 101 was low, went through a maximum at low concentrations of HCOZCH3added to the flow, and decayed rapidly as further HC02CH3was added yielding the cluster ion m / z 133 (eq 14). The data require that k(14) > k(13). F,C-* * HOCH3 HCOZCH, F3C.v (HOCH3)2 + C O ( m / z 101) ( m / z 133) (14) To account for the formation of cluster m / z 63, we suggest that a portion of m / z 133 suffers H + transfer yielding m / z 63 and HCF,. The ratio ( m / z 133)/(m/z 63) = 1 was independent of PHefrom 0.4 to 1.1 torr. Therefore, it appears unlikely that m / z 63 was generated from decomposition of a portion of excited m / z 133 cluster ions since that fraction of m / z 133 decomposing to m / z 63 and HCF, should decrease as collisions with the buffer gas increase. It may be that the secondary Riveros reaction of F3C-- HOCH, with HCO,CH,,produced two isomeric product cluster ions, the stable symmetrically solvated anion 4, and the metastable cluster 5. Cluster 5 might then decompose directly
+
+
-
HOCH3
F
I
4
5
to HCF, and C H 3 0 - - . H O C H 3 ( m / z 63).29 The secondary (25) A kcH3/kcHS = 16 was observed in the reactions of S-.with CFlCO2R which also occured at nearly the collision limit.' (26) Faigle, J. F. G.; Isolani, P. C.; Riveros, J. M. J . Am. Chem. SOC.1976, 98, 2049-2052 and references therein. (27) AHo = +12.7 kcal mol-' for formation of HCF3 + CH,O- + CO in reaction 13." (28) (a) Sullivan, S. A,; Beauchamp, J. L. J . Am. Chem. SOC.1976, 98, 1160-1165. (b) Yamdagni, R.; Kebarle, P. Ibid. 1971, 93, 7139-7143.
J . Am. Chem. SOC., Vol. 105, No. 25, 1983 7211
Nucleophilic Reactions of F 3 C Riveros reaction shown in eq 15 was previously established in the FA, k ( l s )= (1.2 f 0.1) X 10" cm3 molecule-' s-l.'
-
CH30-. HOCH,
+ HC02CH3
-+
CH,O-..(HOCH3),
+ CO
(15)
Reaction of F 3 Cwith C02. The reactions of anions (R-) with CO, yielding carboxylate anion products (RCOY) determined in flow systems are listed as termolecular reactions30 (eq 16) where M is the buffer gas. However, the termolecularity of eq 16 has been established in relatively few examples. R-
+ C02 + M
+
RC02-
+M
(16)
The reaction of F3C- with C 0 2 under our standard operating conditions of PHe= 0.5 torr and D = 80 m s-l produced exclusively the carboxylate anion m / z 113 with a modest apparent bimolecular rate constant, k = (1.2 0.1) X lo-" cm3 molecule-' s-'. Repeating the experiment with the flow exhaust gate valve partially throttled ( P H e= 1.1 torr, 0 = 36 m s-l) gave the same product anion, m / z 113, but the apparent bimolecular rate constant doubled, k = (3.4 f 0.2) X lo-" cm3 molecule-' s-'. This change in rate constant as a function of PHeis outside of maximum errors and indicates that (CF3C02-)*was produced in a distribution of vibrationally excited states which can be stabilized by collisions with third bodies (M) present in the flow (eq 17). Since the
*
F3C-
+ CO, e(F3CC02-)* kgi
kE
kdM1
F3CC0,( m l z 113)
(17)
principal third body (M) present in the flow is He, the termolecular rate constant, k = 1.1 X lo-,' cm6 molecule-* s-', is calculated for this reaction involving helium in collisional stabilization of the initially produced excited product ion. Characterization of the Adduct (CF,),CO- as a Tetrahedral Alkoxide. Brauman et aL31 reported that RCOC12- species, produced by a Cl--transfer ion-molecule reaction to neutral RC(=O)CI, could be minima on the potential surfaces and were "probably best described as a tetrahedral intermediate". Bowie3* and co-workers have observed small signals attributed to adduct ions formed in the reactions of various anions with anhydrides. Our results of the reactions of PhN-. and S-. with ketones also gave small signals of adduct species, the intensity of which increased with increasing PHe.' However, the precise structures of these adducts remain ill defined. The most convincing results of addition adduct formation were described by Bohme et aL9 in the termolecular reactions of H-, HO-, and CH,O- with H2C=0. However, the suggestion that the observed anion products with the mass of the adduct were, in fact, ion-dipole complexes cannot be unequivocally set aside. We, therefore, decided to establish the structure of one of the above adduct ions by determining its proton affinity (PA). Experiments carried out in the FA are unique among gas-phase methods in that ions can be sequentially produced and reacted by the addition of different reactant gases through inlet ports placed along the length of the flow tube. In the design and construction of our present modular flow tube (Figure l), an (29) McDonald, R. N.; Chowdhury, A. K.; Setser, D. W. J . Am. Chem.
addition port was installed 40 cm downstream of the neutral reactant inlet port. If the ion-molecule reaction of interest was fast (to completely remove precursor ions) and a single (or, at least major) product ion resulted, chemical and/or thermochemical information on this product ion can be obtained to verify its structure from the results of further ion-molecule reactions of this ion with other neutral reagents added via this final port. Among the present results, the reaction of F3C- with (CF3)2C=O satisfies both requirements of the original ion-molecule reaction. Since l.iffoadd((cF3),c0H)= PA((CF3)3CO-) = 334.3 f 2 kcal mol-','* potential H+ donors were added through the final port to bracket the PA of the m / z 235 adduct ion. If m j z 235 was the ion-dipole complex (F,C-/(CF,),C=O), we would expect its PA to be close (or equal) to that of F3C- itself. The three potential H+ donors of known gas-phase acidity3a selected for this study were HCO,H (akfoacld= 345.2 f 2 kcal mol-'), HCI (AHoacid = 333.6 f 2 kcal mol-'), and C F 3 C 0 2 H (AHoaad = 322.7 i 2 kcal mol-'). The results of adding separately these three acids to the flow containing only the ion m j z 235 produced by eq 4 are shown in eq 18-20. The fact that no reaction
-
+ HCOzH no reaction m / z 235 + HC1 -.+ Cl- + (CF,),CO-..HCI m / z 235
m / z 235
+ CF3C0,H
CF3C02( m l z 113)
-
( m j z 271, 273)
(18) (19)
+ CF3C02-*.H02CCF3 + CF3COC.. HOC(CF3)3 ( m l z 227)
( m l z 349)
(20) occurred between m / z 235 and H C 0 2 H establishes the structure of the ion m / z 235 as perfluoro-tert-butoxide anion. This is supported by the results from reaction 19 where both products were of similar intensities indicating an equilibrium process between ions and neutrals of similar basicities and acidities, and from reaction 20 where the ions m / z 113 and 227 are the major products and m / z 349 was minor.
Conclusion We consider that the above data are reasonable and sufficient evidence that each of the adducts produced in reactions 3b, 4, 5b, 6a, 7b, 8a, 9a, loa, and 1l b have the corresponding tetrahedral structures formed by nucleophilic addition of F3C- to the carbonyl group of the corresponding neutral reactant. Further, we conclude that this mechanism generally describes the gas-phase reactions of anions with neutral carbonyl-containing molecules and is analogous to that established for related processes in the condensed pha~e.~,.~~ Acknowledgment. We gratefully acknowledge support of this research from the U.S. Army Research Office and the National Science Foundation (equipment grant) and encouragement from Professor D. W. Setser. Registry No. F,C-, 54128-17-5; CH3Br, 74-83-9; CH3CI, 74-87-3; (CH,),C=O, 67-64-1; (CF,)ZC=O, 684-16-2; CH3COZCH3, 79-20-9; C6H,COZCH, 93-58-3; CF3COzCH,,431-47-0; CF,CO,C,H, 383-63-1; CF,C02C(CH,),, 400-52-2; (CH,O)*C=O, 61 6-38-6; HCO2CH3, 10731-3; COZ, 124-38-9.
SOC.1981, 103, 7586-7589.
(30) Albritton (Albritton, D. L. At. Data Nucl. Data Tables 1978, 22, 1-101) lists ion-molecule reaction rate constants measured in flow reactors through 1977. (31) Asubiojo, 0. I.; Blair, L. K.; Brauman, J. I. J . Am. Chem. SOC.1975, 97, 6685-6688. (32) Bowie, J. H. Arc. Chem. Res. 1980, 13, 76-82 and references therein.
(33) Bender, M. L. J . Am. Chem. SOC.1951, 73, 1626-1629; Chem. Reo. 1960, 60, 53-1 13.
(34) For a general discussion and further references, see: March, J. "Advanced Organic Chemistry"; 2nd ed.; McGraw-Hill: New York, 1977; pp 307-31 1.
